In general, speech signal processing involves processing electrical and/or electronic signals for recognition or synthesis of speech. Speech synthesis is the production of speech from text by artificial means, and text-to-speech (TTS) systems provide an alternative to conventional computer-to-human visual output devices like computer monitors or displays. Conversely, automatic speech recognition (ASR) technologies enable microphone-equipped computing devices to interpret speech and thereby provide an alternative to conventional human-to-computer tactile input devices such as keyboards or keypads.
In some environments, TTS and ASR technologies are combined to provide a user with hands-free audible interaction with a system. For example, a telematics system in a vehicle may receive text messages, use TTS technology to present them in audible form for a driver, receive a verbal reply from the driver, and relay the reply to a server which recognizes the reply and generates and transmits a corresponding text message response. Speech recognition is normally difficult task, but especially so when dealing with strange abbreviations and other colloquial peculiarities of modern day text messaging.